Analyze-String and A Multiple-Match Example

mphare · January 27th, 2009, 04:54 PM

I'm using Saxon 9b

I using XSLT 2.0 3rd ed as reference.

I have a string $keyVal = "<ABC>,<XYZ>" and I need to parse out the ABC and the XYZ. Note, that at run time I will not know how many comma separated <AAA> sets there will be. They will not all have 3 characters and they may contain Letters, Numbers and a Dash symbol. So, "<IB6F-3-4>,<IB6S-23-19>,<IFA2-3>" would be a perfectly general example. I would expect IB6F-3-4 and IB6S-23-19 and IFA2-3 to be the parsed results.

I have tried

<xsl:analyze-string select="$keyVal" regex="^.*?([A-Z0-9\-]+).*?$">
<xsl:matching-substring>
<xsl:value-of select="regex-group(1)"/>
</xsl:matching-substring>
</xsl:analyze-string>

Which very nicely parses the first item, but not the rest.

I'm reading the section in Chpt 5 called "A Multiple-Match Example" which I believe is what I need, but I'm not understanding what I'm reading. It seems that my regex should have called matching-substring for each pattern that matched Letters, Numbers and/or a Dash. But I'm only getting it called once.

What part am I missing?

Thanks.

mhkay · January 27th, 2009, 05:05 PM

Your regex is anchored to the start and end of the string by virtue of the ^...$ symbols, which means the whole string either matches (once) or non-matches (once). To do a multiple match, remove the anchors and the internal repetition so it becomes regex="[A-Z0-9\-]+", and you will then find that <xsl:matching-substring> is invoked once for each maximal sequence of these characters. You then don't need to use regex-group(), you can access the matched substring as ".".

mphare · January 27th, 2009, 05:21 PM

Yes, I see. That was my misunderstanding of how the regex worked. I've always been used to anchoring everything in other languages..

But that worked like a champ!

Thanks very much!